differently observed, it has to be determined how great the limits of error are, on the average. It has been found, that most observers are capable of reading with accuracy to 0.02 of a degree. In order to attain greater certainty, we shall consider even a difference of 0.04 of a degree as the limit of error. Only with a change of rotation of 0.05 of a degree can decomposition be assumed to have taken place. The limit can thus be fixed without any danger, because, when an hydrolysis of the peptone does take place, the alteration of rotation is certainly more than 0.04 of a degree.
This method, as such, presents hardly any sources of error. At most, errors may be occasionally produced through turbidity, precipitates, and the like. Fortunately, however, in such cases, which actually very seldom happen with proper working, the reading of the rotation is impossible, and so this source of error disappears of itself. Of course, we should get no result if we were to try to polarize a cloudy solution.
A very important source of error would arise, if the range of rotation of the cold solution were taken for the initial value. The readings must be taken from the moment the contents of the tubes reach a temperature of 37° C. It is best to take the reading at the end of one hour, and take another at the end of the second hour. Values obtained in such a manner should, in general, not be too distant one
